Generating accomplishments

PDS Accomplishment Status Report Generator

Generates a canonical PDS EN accomplishment report from a activity.json file, grouped by PDS product team. Can optionally run the pds-activity CLI (from the lasso-issues tool) to collect the data first.

Inputs

• activity.json path (optional): Path to an existing activity.json file. If not provided, the skill will offer to run pds-activity to generate one.
• Clarifying answers (gathered interactively before generating): reporting period label, audience, priority themes, scope, format.

Prerequisites

• If running pds-activity to collect data: Python 3.10+, lasso-issues installed (see below), GITHUB_TOKEN env var set
• GitHub token: export GITHUB_TOKEN=<your-token> (needs repo read scope for NASA-PDS org)

---

Step 0 — Check pds-products.yaml for upstream changes

Before doing anything else, check whether pds-products.yaml (in this skill's directory) matches the upstream canonical version.

Locate the skill directory: pds-products.yaml lives alongside this SKILL.md — typically at ~/.claude/plugins/cache/pds-agent-skills/creating-pds-issues/<hash>/skills/generating-accomplishments/pds-products.yaml.

Fetch the upstream and diff:

curl -fsSL https://raw.githubusercontent.com/NASA-PDS/.github/main/conf/pds-products.yaml > /tmp/pds-products-upstream.yaml
diff /tmp/pds-products-upstream.yaml <path-to-local-pds-products.yaml>

If the diff shows changes to products: entries (new repos, new products, changed ignore: or work_stream: or core_backbone: values):

1. Apply those structural changes to the local copy — preserve our local comment additions (the reporting-exclusion notes at the top).
2. Do not copy upstream version: or updated: fields — those are intentionally omitted from our copy.
3. Confirm: "Updated pds-products.yaml — applied N changes from upstream."

If the fetch fails (network error, non-200): warn and continue with the local copy.

⚠️ Could not check pds-products.yaml for updates (fetch failed). Using local copy. Check manually: https://github.com/NASA-PDS/.github/blob/main/conf/pds-products.yaml

If diff shows only comment/whitespace differences: proceed silently — no update needed.

---

Step 1 — Determine Input Source

Check whether an activity.json was provided by the user (as a file path argument or mentioned in conversation).

If provided: load it directly. Confirm: "I found <path> — using that as the data source."

If NOT provided: ask the user:

I don't see an activity.json. I can either:

1. Run pds-activity to collect fresh data from GitHub (requires GITHUB_TOKEN and lasso-issues installed)
2. Use a file you point me to

Which would you prefer, and if option 1, what date range should I collect? (e.g., 2026-06-01 to 2026-06-30)

Then proceed per their answer:

• Option 1: Follow Step 1a — Run pds-activity below.
• Option 2: Ask for the file path and load it.

Step 1a — Run pds-activity

Verify pds-activity is available:

pds-activity --version 2>/dev/null || python -m lasso.issues.activity.cli --version 2>/dev/null

If missing, offer to install from main branch:

pip install git+https://github.com/NASA-PDS/lasso-issues.git@main

Or if the repo is checked out locally at ~/proj/pds/pdsen/workspace/lasso-issues:

pip install -e ~/proj/pds/pdsen/workspace/lasso-issues

Once available, run collection — save output to ~/.claude/tmp/activity.json:

mkdir -p ~/.claude/tmp
pds-activity \
  --org NASA-PDS \
  --start-date <START> \
  --end-date <END> \
  --output ~/.claude/tmp/activity.json

Confirm success: "Collected activity data — <N> issues, <M> PRs, <K> releases."

---

Step 2 — Ask Clarifying Questions

Before generating the report, ask the user ALL of the following questions in a single message. If the user says "just go" or "use defaults", apply the defaults shown and skip ahead.

Before I generate the report, a few quick questions:

1. **Reporting period label** — What should I call this period?
   (e.g., "Sprint 142", "Q3 FY2026", "June 2026", "FY2026 Annual")
   Default: derived from the activity.json date range

2. **Priority themes or highlights** — Are there any areas you want foregrounded or called out?
   (e.g., "Highlight Registry work", "Focus on cloud migration", "Lead with release milestones")
   Default: none — use natural grouping

4. **Scope** — Which product groups should I include?
   Options: All groups / Specific groups (list them) / Exclude specific groups
   Default: all groups with activity

5. **Output format** — Any specific format requirements?
   (e.g., Markdown, slide bullet points, email summary, Confluence wiki markup)
   Default: Markdown

Wait for answers before proceeding to Step 3.

---

Step 3 — Parse and Group Activity Data

Load the activity.json. The top-level structure is:

{
  "metadata": { "org": "...", "start_date": "...", "end_date": "...", "repo_count": N, ... },
  "issues": [ { "repo": "...", "title": "...", "state": "...", "labels": [...], "closed_at": "...", "html_url": "...", ... } ],
  "pull_requests": [ { "repo": "...", "title": "...", "merged_at": "...", "html_url": "...", ... } ],
  "releases": [ { "repo": "...", "tag": "...", "name": "...", "published_at": "...", "html_url": "...", ... } ]
}

3a — Build the repo → product group map

Use the product groupings embedded in pds-products.yaml in this skill directory. Map each repo name to its product group name. Repos not found in any product's repositories list go into Uncategorized.

Products with ignore: true (dependencies, node-products, archived_repositories) should be silently excluded — do not surface them in the report.

Work stream grouping (use as top-level sections in this order):

1. core-data-services — Core Data Services
2. planetary-data-cloud — Planetary Data Cloud
3. web-modernization — Web Modernization
4. (no work_stream) — Other / Cross-cutting

Within each work stream, group by product name (use the product's description field as a human-readable label).

Core backbone products (core_backbone: true) should be visually distinguished — add a ⭐ suffix to their section header.

3b — Collect accomplishments per product group

For each product group, gather:

Closed issues — issues where state == "closed" and closed_at is within the reporting window. Each issue is one accomplishment bullet.

Merged PRs — pull_requests where merged_at is within the reporting window. Surface only if NOT already represented by a linked issue (check linked_issues field — if non-empty and those issues are already captured, skip the PR to avoid duplication). If a PR has no linked issue, surface it as a standalone accomplishment.

Releases — releases where published_at is within the reporting window. Always surface releases as accomplishments (they represent completed milestones).

Filter out noise: Skip issues/PRs with labels: duplicate, invalid, wontfix, icebox, spam. Skip bot-authored PRs (author contains [bot]).

Exclude ignored products: Any repo belonging to a product with ignore: true in pds-products.yaml (ops, dependencies, node-products, archived_repositories) must be silently excluded from all counts, bullets, and the metrics section. Never surface these repos anywhere in the report.

3c — Compute metrics

Before generating the report, compute the following counts across all non-ignored repos.

Issue type counts — classify each closed issue by its labels using this priority order (first match wins):

Label(s)	Type
bug, defect	Bug
requirement, feature	Requirement / Feature
enhancement, improvement	Enhancement
task	Task
theme	Theme
security	Security
breaking-change, backwards-incompatible	Breaking Change
(none of the above)	Other

Track per-type counts and per-product-group breakdowns. Also track:

• Releases shipped: count of releases items in the window
• PRs merged (standalone, not already counted via linked issue): count
• Repositories active: count of distinct repos with at least one closed issue, merged PR, or release
• Themes addressed: count of distinct parent issues / themes referenced by any closed issue
• Must-have items completed: count of closed issues with p.must-have label
• Should-have items completed: count of closed issues with p.should-have label

3d — Detect highlights

Across ALL product groups, identify the top 5–8 accomplishments that represent the most significant work. Criteria:

• Issues linked to a theme or p.must-have label
• Issues whose parent issue is a release theme
• Releases (always significant)
• Issues/PRs with labels: breaking-change, backwards-incompatible, security, requirement
• Issues with closing_release set (delivered in a release)

---

Step 4 — Generate the Report

Report structure

# PDS EN Accomplishments — <REPORTING_PERIOD>

> **Period:** <START_DATE> – <END_DATE>

---

## At a Glance

| Metric | Count |
|--------|-------|
| 🐛 Bugs resolved | N |
| ✨ Requirements / Features delivered | N |
| 🔧 Enhancements completed | N |
| ✅ Tasks completed | N |
| 🎯 Themes addressed | N |
| 🔒 Security issues resolved | N |
| ⚠️ Breaking changes | N |
| 🚀 Releases shipped | N |
| 📦 Repositories active | N |
| 🔀 PRs merged (standalone) | N |
| 🏆 Must-have items completed | N |
| 📋 Should-have items completed | N |

*(Excludes noise labels: duplicate, invalid, wontfix, icebox, spam)*

---

## Highlights

- <Top accomplishment 1> ([repo#N](url))
- <Top accomplishment 2> ...
(5–8 bullets, outcome-focused, one link per bullet)

---

## Core Data Services

### <Product Name> [⭐ if core_backbone]
*<product description>*

**Releases**
- `v<tag>` — <release name> ([<repo>](url))

**Completed Work**
- <Issue/PR title> ([<repo>#<N>](url))
- ...

(repeat for each product with activity in this work stream)

---

## Planetary Data Cloud

(same structure)

---

## Web Modernization

(same structure)

---

## Uncategorized

*Repositories not mapped to a product group:*

- **<repo-name>**: <title> ([#N](url))
- ...

---

## Summary by Work Stream

| Work Stream | Products Active | Bugs | Requirements | Enhancements | Tasks | Releases |
|-------------|----------------|------|--------------|--------------|-------|----------|
| Core Data Services | N | N | N | N | N | N |
| Planetary Data Cloud | N | N | N | N | N | N |
| Web Modernization | N | N | N | N | N | N |
| Uncategorized | — | N | N | N | N | N |
| **Total** | **N** | **N** | **N** | **N** | **N** | **N** |

Bullet style rules

• Start with a verb: "Adds…", "Fixes…", "Delivers…", "Implements…", "Resolves…"
• Keep bullets ≤ ~15 words before the link
• Every bullet MUST end with a GitHub link in the form ([repo#N](url)) or ([repo](url)) for releases
• Breaking changes get a ⚠️ prefix: ⚠️ Breaking: removes deprecated API v1 ([registry#123](url))
• Security fixes get a 🔒 prefix
• For releases, format as: `v1.4.0` released — <brief description> ([registry](url))
• If a parent issue / theme is known, add it as a sub-bullet: ↳ Theme: [Sprint 141 Theme: Cloud Migration](url)
• Omit product sections with zero activity (don't create empty headers)

---

Step 5 — Present and Offer Export

Display the full report in the conversation.

Then ask:

Would you like me to:

1. Save this to a file (e.g., accomplishments-<period>.md)?
2. Adjust any sections, reorder priorities, or change the format?
3. Generate a condensed version (e.g., slide bullets or email summary)?

Apply any requested changes and save if asked.

---

Edge Cases

• Empty product groups: Silently omit from report — do not create headers with no content.
• Repo not in products.yaml: Goes to Uncategorized. Never fail on unknown repos.
• Duplicate accomplishments (same issue appears in both issues and linked from a PR): Deduplicate — show the issue once, skip the PR.
• No activity in a work stream: Omit that entire work stream section.
• Releases with empty body_summary: Show as `vtag` released with just the link.
• Very large datasets (500+ items): Summarize by product group rather than listing every item — cap each product at 10 bullets, add "(and N more)" if truncated.
• Missing dates (closed_at or merged_at is null): Include the item but note "(date unknown)".
• Parent issue / theme linking: If parent_issue is present on an ActivityIssue, show the theme title as a sub-bullet to provide context.

---

Additional Resources

• pds-products.yaml — PDS product-to-repository mapping used for grouping
• lasso-issues CLI docs — source for pds-activity tool